New task aligns molecular structure with gene expression changes.
problem Modeling the relationship between chemical structure and gene expression changes.
method Developed a cross-modal small molecule retrieval task and a coordinated deep learning approach to align chemical structure and gene expression profiles.
result Demonstrated the feasibility of the new task and highlighted the limitations of current data and systems.
Tabular in-context learners perform well on biomolecular tasks, but performance depends on the representation used.
problem Predicting biomolecular properties from limited labeled data.
method Evaluating tabular in-context learners on protein fitness regression and small-molecule classification tasks.
result Tabular in-context learners are competitive for protein fitness regression but not for small-molecule classification.
SparseChem speeds up ML for small molecules.
problem Training fast and accurate ML models for high-dimensional data.
method Supports millions of features and compounds, trains various models.
result Fast and accurate machine learning models for biochemical applications.
Neural net reweighing improves selectivity in molecule binding studies.
problem Improving selectivity in neural net models for molecule binding studies.
method Greedy algorithm to reweight loss function based on Wasserstein distance.
result Proven to make neural net weights approach limiting distribution of another dataset.
A neural network and evolutionary algorithm framework designs nonlinear optical molecules.
problem Designing efficient nonlinear optical materials.
method Multi-stage Bayesian neural network (msBNN) and corrected Lewis-mode group contribution method (cLGC) combined with evolutionary algorithm (EA).
result Accurately and efficiently designs molecules with different optical properties using a small data set.
LaMBO optimizes biological sequences using autoencoders and Bayesian optimization.
problem Bayesian optimization for drug design is limited by discrete, high-dimensional decision variables.
method Jointly trains denoising autoencoder with a Gaussian process head for gradient-based optimization in latent space.
result LaMBO outperforms genetic optimizers and requires no large pretraining corpus.
New method refines model predictions as design evolves.
problem Designing objects with desired properties using data-driven methods.
method Formalized as a game, developed autofocusing strategy for model retraining.
result Autofocusing improves model predictions in design space.
Autodock is a widely used molecular modeling tool which predicts how small molecules bind to a receptor of known 3D structure. The current version of AutoDock uses meta-heuristic algorithms in combination with local search methods for doing the conformation search. Appropriate settings of hyperparameters in these algor…
DESMILES uses deep learning to improve drug discovery by optimizing molecule properties.
problem Improving the efficiency and accuracy of drug discovery through better molecular design.
method DESMILES is a deep neural network model that optimizes molecular properties for drug discovery.
result DESMILES achieved a 77% lower failure rate in modifying molecules to inhibit the dopamine receptor D2 compared to state-of-the-art models.
AI and HPC help screen millions of molecules for SARS-CoV-2 treatments.
problem Finding effective treatments for SARS-CoV-2.
method AI and HPC enable screening of large molecule datasets.
result Data release of 23 datasets with 4.2 billion molecules.
Recent advances in machine learning have made significant contributions to drug discovery. Deep neural networks in particular have been demonstrated to provide significant boosts in predictive power when inferring the properties and activities of small-molecule compounds. However, the applicability of these techniques …
When confronted with a substance of unknown identity, researchers often perform mass spectrometry on the sample and compare the observed spectrum to a library of previously-collected spectra to identify the molecule. While popular, this approach will fail to identify molecules that are not in the existing library. In r…
Automates GNN design for molecular property prediction.
problem Designing and tuning GNN architectures for molecular property prediction is labor-intensive.
method Developed a NAS approach to automatically discover high-performing GNN architectures for MPNNs.
result Automatically discovered MPNNs outperform manually designed GNNs in molecular property prediction.
Researchers parallelize neural kernels for large-scale data, achieving state-of-the-art accuracy.
problem Limited scalability of neural kernels on large datasets.
method Massively parallel computation across many GPUs, combined with a distributed, preconditioned conjugate gradients algorithm.
result Achieved state-of-the-art accuracy of 91.2% on CIFAR-5m dataset using neural kernels.
Generative model designs drug combinations for improved efficacy and reduced side effects.
problem Designing effective drug combinations to overcome resistance and reduce side effects.
method Developed a deep generative model using HVGAE and a novel reward system.
result Network-principled drug combinations show reduced toxicity and potential for new strategies.
A deep neural network based architecture was constructed to predict amino acid side chain conformation with unprecedented accuracy. Amino acid side chain conformation prediction is essential for protein homology modeling and protein design. Current widely-adopted methods use physics-based energy functions to evaluate s…
CogMol designs novel drug-like molecules for SARS-CoV-2 targets.
problem Designing efficient drugs for novel viral proteins.
method End-to-end framework combining VAE, controlled sampling, and predictors.
result Highly selective and affinity molecules for SARS-CoV-2 targets.
SILVR generates new molecules fitting protein binding sites.
problem Generating novel small molecule compounds for drug design.
method Selective Iterative Latent Variable Refinement (SILVR) for diffusion-based molecule generation.
result SILVR can generate new molecules similar in shape to original fragments without protein knowledge.
Deep generative model discovers inhibitors for unknown targets.
problem Discovering novel inhibitor molecules for unknown drug targets.
method Deep generative framework trained on protein sequences, small molecules, and interactions.
result Micromolar-level inhibition observed for two out of four synthesized candidates, including activity against SARS-CoV-2 variants.
A new training method for efficient Boltzmann generators.
problem Training equivariant continuous normalizing flows (CNFs) is computationally expensive.
method Equivariant flow matching, based on optimal transport flow matching.
result Equivariant flow matching yields more efficient flows with shorter integration paths.
Deep convolutional neural networks comprise a subclass of deep neural networks (DNN) with a constrained architecture that leverages the spatial and temporal structure of the domain they model. Convolutional networks achieve the best predictive performance in areas such as speech and image recognition by hierarchically …
We introduce a deep learning architecture for structure-based virtual screening that generates fixed-sized fingerprints of proteins and small molecules by applying learnable atom convolution and softmax operations to each compound separately. These fingerprints are further transformed non-linearly, their inner-product …
This abstract reviews recent methods for predicting protein-ligand binding affinity.
problem Predicting protein-ligand binding affinity for various applications in life sciences.
method Traditional and deep learning models for binding affinity prediction.
result Improved predictive performance of AI-driven models.
Molecular "fingerprints" encoding structural information are the workhorse of cheminformatics and machine learning in drug discovery applications. However, fingerprint representations necessarily emphasize particular aspects of the molecular structure while ignoring others, rather than allowing the model to make data-d…
In this study, we intend to solve a mutual information problem in interacting molecules of any type, such as proteins, nucleic acids, and small molecules. Using machine learning techniques, we accurately predict pairwise interactions, which can be of medical and biological importance. Graphs are are useful in this prob…
CoDrug uses KDE to create valid prediction sets for drug molecules under covariate shift.
problem Creating reliable uncertainty estimates for drug properties from computational models.
method CoDrug employs an energy-based model and KDE to assess and rectify distribution shift.
result CoDrug reduces the coverage gap by over 35% compared to non-adjusted conformal prediction sets.
A molecule's geometry, also known as conformation, is one of a molecule's most important properties, determining the reactions it participates in, the bonds it forms, and the interactions it has with other molecules. Conventional conformation generation methods minimize hand-designed molecular force field energy functi…
Branching Flows generates sequences of varying lengths using binary trees.
problem Generating sequences of unknown lengths or fixed elements.
method A generative modeling framework that evolves states over binary trees, controlling sequence length.
result Branching Flows can generate sequences of varying lengths and mix different types of state spaces.
XIMP improves molecular property prediction by integrating multiple graph representations.
problem Graph neural networks struggle in data-scarce regimes and fail to surpass traditional methods.
method Cross-graph inter-message passing with multiple graph abstractions.
result XIMP outperforms state-of-the-art baselines across diverse molecular property tasks.
Predicating macroscopic influences of drugs on human body, like efficacy and toxicity, is a central problem of small-molecule based drug discovery. Molecules can be represented as an undirected graph, and we can utilize graph convolution networks to predication molecular properties. However, graph convolutional network…
PUMA interprets metabolomics data to predict pathway activity and assign chemical identities.
problem Interpreting metabolomics data to determine biochemical pathway activities.
method Generative probabilistic modeling using stochastic sampling.
result PUMA predicts pathway activity and assigns chemical identities to metabolites.
Graph neural networks improve odor prediction from molecular structure.
problem Predicting odor from molecular structure is challenging and important.
method Used graph neural networks for QSOR modeling.
result Graph neural networks significantly outperform prior methods on a novel data set.
Predicting bioactivity and physical properties of small molecules is a central challenge in drug discovery. Deep learning is becoming the method of choice but studies to date focus on mean accuracy as the main metric. However, to replace costly and mission-critical experiments by models, a high mean accuracy is not eno…
POEM predicts drug properties without tuning, outperforming other methods.
problem Predicting drug properties from molecular structures efficiently.
method POEM combines multiple molecular representations without hyperparameter tuning.
result POEM outperforms industry-standard methods across 17 tasks.
MACE architecture outperforms alternatives in various molecular and materials science tasks.
problem Improving machine learning force fields for diverse molecular and materials science applications.
method Evaluation of MACE architecture on various datasets and tasks, demonstrating data efficiency and excellent performance.
result MACE architecture generally outperforms alternatives across a wide range of systems, including amorphous carbon, universal materials modelling, and organic chemistry.
Deep neural network identifies potential SARS-CoV-2 inhibitors.
problem Finding novel therapies for SARS-CoV-2.
method Used ChemAI, a deep neural network trained on 220M data points, to screen and rank one billion molecules from the ZINC database.
result Identified 30,000 top-ranked compounds for further bioassays.
Method learns molecular Hamiltonian for accurate electron dynamics predictions.
problem Predict electron dynamics in molecules using learned Hamiltonians.
method Combines linear statistical model with quantum Liouville equation time discretization.
result Predicted electron dynamics closely matches ground truth, even beyond training data.
Enhances neural networks with prior function values to improve accuracy.
problem Improving neural network accuracy in regions without training data.
method Develops a probabilistic approach to augment BNNs with prior function values.
result Predictions rely more on prior information in uncertain regions.
New model simplifies symmetry handling in generative AI.
problem Symmetry handling in generative models for scientific tasks.
method Quotient-space diffusion models, viewing symmetry as quotient space.
result Improves performance over existing methods for molecular structure generation.
This work improves online fine-tuning of diffusion models for specific properties.
problem Efficiently fine-tuning diffusion models to maximize specific properties.
method A novel reinforcement learning procedure that efficiently explores feasible samples.
result The method provides a regret guarantee and empirical validation across multiple domains.
A deep probabilistic model analyzes DNA-encoded library data for efficient screening.
problem Complex data from DNA-encoded library experiments mask underlying signals.
method Compositional deep probabilistic model of DEL data, modeling latent reactions between synthons.
result DEL-Compose model demonstrates strong performance and valuable insights.
With the rapid development of high-throughput technologies, parallel acquisition of large-scale drug-informatics data provides huge opportunities to improve pharmaceutical research and development. One significant application is the purpose prediction of small molecule compounds, aiming to specify therapeutic propertie…
Predicting the biological function of molecules, be it proteins or drug-like compounds, from their atomic structure is an important and long-standing problem. Function is dictated by structure, since it is by spatial interactions that molecules interact with each other, both in terms of steric complementarity, as well …
New models suggest molecules that are often unfeasible to synthesize.
problem Models suggest molecules that are difficult to synthesize.
method Used a computer-aided synthesis planning program to analyze synthesizability of molecules generated by state-of-the-art models.
result State-of-the-art models generate molecules that are often unfeasible to synthesize.
Semi-supervised learning improves QSAR model predictions for novel compounds.
problem Improving model predictions for compounds not in the training set and adjusting for selection bias.
method Semi-supervised learning framework to estimate model quality and adjust for selection bias.
result Predictions for novel compounds are improved by accounting for compound similarity and selection bias.
Background: Pharmacokinetic evaluation is one of the key processes in drug discovery and development. However, current absorption, distribution, metabolism, excretion prediction models still have limited accuracy. Aim: This study aims to construct an integrated transfer learning and multitask learning approach for deve…
Wavelet scattering predicts material properties beyond training data.
problem Predict material properties beyond training data.
method Atomic orbital wavelet scattering transform.
result Extrapolation of material properties achieved.
Machine learning has emerged as an invaluable tool in many research areas. In the present work, we harness this power to predict highly accurate molecular infrared spectra with unprecedented computational efficiency. To account for vibrational anharmonic and dynamical effects -- typically neglected by conventional quan…